Optical communication has many benefits over transmission of signals in an electrical domain. For one, the losses in an optical medium are much less than those incurred in an electrical medium. As a result, signals can travel greater distances through an optical fiber before any necessary regeneration. Another advantage is that optical signals are much less susceptible to electromagnetic radiation. An electrical medium, such as a coaxial cable, generates electromagnetic fields as the signals travel down the cable. These electromagnetic fields can induce noise in neighboring cables and cause interference with the signals traveling on such cables. In addition, noise can be induced upon the coaxial cable signals from the electromagnetic fields generated by the neighboring cables. For these and many other reasons, optical communication is often the preferred mode of communication.
In optical communication networks, as with electrical communication networks, the network needs some manner of adding and dropping signals at points throughout the network. One approach to adding and dropping signals onto an optical medium involves the use of regeneration devices. These regeneration devices convert the optical signals traveling along the optical medium into the electrical domain and route these electrical signals to terminal equipment. Any electrical signals that need to be added and forwarded on to the optical medium are added to the electrical signals that had been detected. A combination of these signals is then converted into optical signals and passed along the optical medium. Some drawbacks to this approach include the loss of signal quality in needing to convert the optical signals into electrical signals and then back to optical signals at each node or station throughout the network, the accompanying loss of speed and increase in latency, and the limitations in bandwidth associated with the electrical medium.
Rather than coupling signals in the electrical domain, a preferred device for coupling signals operates purely in the optical domain. By maintaining the signals in the optical domain, the coupling device can maintain signal quality, operate at higher speeds and at an increased bandwidth. U.S. Pat. No. 5,901,260, which is incorporated by reference, is an example of an optical interface device operating solely in the optical domain. This optical interface device is useful in routing optical signals traveling along either direction on an optical bus to a node and for directing signals from that node onto the bus in both directions. With such an optical interface device, signals that originate at any node within a network can be transmitted to every other node and, conversely, signals from all of the nodes are received at each node. This type of optical interface device is useful in an optical transport system described in U.S. Pat. No. 5,898,801, which is incorporated by reference.
While optical interface devices have many advantages over electrical interface devices, optical interface devices can still limit the performance of the network. For example, each time optical signals are diverted off of an optical bus to a node, the optical interface device necessarily reduces the optical signal level. Consequently, after a certain number of nodes, the remaining signal has such a low optical signal to noise level that the signal is underneath the noise floor and can no longer be detected. In addition to those losses due to splitting of the signal at each node, the optical interface device also introduces losses resulting from the imperfect coupling of light from an input optical fiber to the optical interface device, from the optical interface device to an exit optical fiber, and from the optical interface device to the terminal equipment. The optical interface device therefore introduces losses at each node, which causes the signal to vary at points throughout the network.
U.S. Pat. No. 5,898,801 describes a network having a number of optical interface devices that bi-directionally amplifies the optical signals. By amplifying the optical signals traveling along the optical bus, the number of nodes along the network can be greatly increased. The optical amplifier may comprise a fiber amplifier and, more particularly a rare earth doped fiber amplifier. The doped fiber amplifier is energized with an excitation light, typically at 980 nanometers. This fiber amplifier may be located between nodes along the bus and/or between the terminal equipment and the optical interface device.
As described in U.S. Pat. No. 5,898,801, the length of the rare earth doped fiber influences the amount of amplification provided by the fiber amplifier. By placing the fiber amplifiers between each node, the fiber amplifiers can compensate for losses incurred by splitting the signals at each node. Thus, a signal that originates at one end of the bus can travel along the bus, have a fraction of the signal diverted at each node, and then be amplified after incurring those losses. This approach to amplification, as mentioned above, greatly increases the number of nodes that may be in a network. This approach to amplification, however, is more challenging when the network topology is dynamic. For example, a network may have different amplification needs with an initial set of nodes than it would need later when nodes are added at other points within the network, are removed from the network, or are placed at different points within the network. The placement of fiber amplifiers at certain points may therefore not be optimal for a different configuration of nodes on the network. As a result, the signal level and quality of the signal may vary throughout the network. For certain types of signals, these variations may not affect performance of the network overall. On the other hand, for other types of signals, such as radio frequency (RF) signals and other analog signals, maintaining a consistent signal dynamic range and waveform quality throughout the network is imperative.